We build the training data and benchmarks that frontier AI labs use to make their models better at hardware engineering. The artifacts you produce become the ground truth these models learn from and are measured against.
In this track you will develop, maintain, and improve the pipelines and automation that let us produce RTL design and verification tasks at scale. This is fundamentally a software role with hardware literacy: you build the tooling that the design and verification engineers depend on every day.
You will also work LLM-in-the-loop every day: using the newest frontier models as tools while producing the very data that sharpens them. A large part of the job is reasoning about where these models succeed, fail, and can be gamed.
Your Mission
Build the pipelines: Develop and maintain the automation and workflows used to produce, review, and ship tasks at scale.
Orchestrate LLMs: Integrate frontier model APIs into task-generation and quality workflows.
Improve throughput and quality: Instrument the pipeline, remove bottlenecks, and raise the reliability of what ships.
Uphold rigor: Maintain the highest standards of coding, debugging, and documentation across every deliverable.
Core Requirements
Non-negotiable these are what we screen for.
Education:
BS/MS in a technical field, or equivalent practical experience.
Experience: 2+ years building production software and automation (Python or similar).
Integration: API integration and pipeline / workflow tooling.
Containerization: Comfortable building, debugging, and operating Docker-based environments -- images, resource limits, orchestrating many containers at once. This is how every task runs.
RTL literacy: Enough hardware / RTL literacy to build tooling around the design and verification lifecycle.
LLM tooling: Comfort using state-of-the-art LLMs as daily working tools.
Communication: Exceptional written communication much of the output is documentation.
Ownership: You close loops without being chased.
Helpful, But Well Teach You
Not required. This field is current — most of it can be learned on the job, and a robust engineer picks it up quick.
LLM orchestration: Experience wiring LLM APIs into real production workflows.
Workflow platforms: Asana / Slack APIs, Apps Script, or comparable automation surfaces.
LLM training data: Prior experience with training-data, benchmark, or eval pipelines.
Hardware background: Any prior RTL design or verification exposure.